The MAV_TYPE enum was not in sync with the mavlink specs anymore. It
makes therefore sense to remove the duplication and include the correct
mavlink header file where it is needed.
Also, error counts which are not populated, can be scrapped.
Since the circuit breaker bools are not actually used anywhere else than
in the commander, it is safe to remove them and replace them with local
bools.
this causes only problems and is not supported by ANSI C. error caused:
fatal error: field '_buf' with variable sized type 'ubx_buf_t' not at the end of a struct or class is a GNU extension
the warning is:
error: ignoring return value of ‘ssize_t write(int, const void*, size_t)’, declared with attribute warn_unused_result [-Werror=unused-result]
- Note that the simulator still uses gpssim by default
- now the gps module can be used in the SITL. this makes it possible to test
the real gps HW under POSIX
additional steps needed to use it:
- in the rcS_jmavsim_iris, make sure to start the gps instead of gpssim:
gps start -d /dev/ttyACM0 -s
- disable the mavlink serial connection in simulator_mavlink.cpp,
openUart(PIXHAWK_DEVICE, 115200);
- this also fixes a memory leak in the gps module
This should help the Snapdragon to connect if the network is not yet up
by the time the mainapp starts. It will retry to find a network and
broadcast once it finds one.
The timestamp in the DriverFramework is not the same as the one
provided by hrt_absolute_time() in PX4. Because we are publishing on the
PX4 side, let's also use the timestamp from there.
This adds a call on startup of the muorb on the aDSP side to use an
offset for hrt_absolute_call(). This means that the hrt_absolute_call()
on the appsproc (Linux) side should now match the one on the aDSP
(QURT) side.
The accuracy still needs to be determined.
Julian Oes